RESEARCH ARTICLE

A cluster of Ankyrin and Ankyrin-TPR repeat

genes is associated with panicle branching

diversity in rice

Giang Ngan KhongID1*, Nhu Thi LeID

1, Mai Thi PhamID1, Helene AdamID

2, Carole Gauron2,

Hoa Quang LeID3, Dung Tien PhamID

3, Kelly ColongesID1¤, Xuan Hoi PhamID

1, Vinh

Nang Do1, Michel Lebrun1,4, Stefan JouannicID1,2*

1 LMI RICE, National Key Laboratory for Plant Cell Biotechnology, Agronomical Genetics Institute, Hanoi,

Vietnam, 2 UMR DIADE, University of Montpellier, IRD, Montpellier, France, 3 School of Biotechnology and

Food Technology, Hanoi University of Science and Technology, Hanoi, Vietnam, 4 UMR LSTM, University of

Montpellier, IRD, CIRAD, INRAE, SupAgro, Montpellier, France

¤ Current address: CIRAD, UMR AGAP, University of Montpellier, INRAE, Montpellier, France

* ngangiang.khong2010@gmail.com (GNK); stephane.jouannic@ird.fr (SJ)

Abstract

The number of grains per panicle is an important yield-related trait in cereals which depends

in part on panicle branching complexity. One component of this complexity is the number of

secondary branches per panicle. Previously, a GWAS site associated with secondary

branch and spikelet numbers per panicle in rice was identified. Here we combined gene

capture, bi-parental genetic population analysis, expression profiling and transgenic

approaches in order to investigate the functional significance of a cluster of 6 ANK and

ANK-TPR genes within the QTL. Four of the ANK and ANK-TPR genes present a differential

expression associated with panicle secondary branch number in contrasted accessions.

These differential expression patterns correlate in the different alleles of these genes with

specific deletions of potential cis-regulatory sequences in their promoters. Two of these

genes were confirmed through functional analysis as playing a role in the control of panicle

architecture. Our findings indicate that secondary branching diversity in the rice panicle is

governed in part by differentially expressed genes within this cluster encoding ANK and

ANK-TPR domain proteins that may act as positive or negative regulators of panicle meri-

stem’s identity transition from indeterminate to determinate state.

Author summary

Grain yield is one of the most important indexes in rice breeding, which is controlled in

part by panicle branching complexity. A new QTL with co-location of spikelet number

(SpN) and secondary branch number (SBN) traits was identified by genome-wide associa-

tion study in a Vietnamese rice landrace panel. A set of four Ankyrin and Tetratricopep-

tide repeat domain-encoding genes was identified from this QTL based on their

difference of expression levels between two contrasted haplotypes for the SpN and SBN

traits. The differential expression is correlated with deletions in the promoter regions of

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 1 / 24

a1111111111

a1111111111

a1111111111

a1111111111

a1111111111

OPEN ACCESS

Citation: Khong GN, Le NT, Pham MT, Adam H,

Gauron C, Le HQ, et al. (2021) A cluster of Ankyrin

and Ankyrin-TPR repeat genes is associated with

panicle branching diversity in rice. PLoS Genet

17(6): e1009594. https://doi.org/10.1371/journal.

pgen.1009594

Editor: Sarah Hake, "USDA-ARS Pacific West Area",

UNITED STATES

Received: March 8, 2021

Accepted: May 10, 2021

Published: June 7, 2021

Peer Review History: PLOS recognizes the

benefits of transparency in the peer review

process; therefore, we enable the publication of

all of the content of peer review and author

responses alongside final, published articles. The

editorial history of this article is available here:

https://doi.org/10.1371/journal.pgen.1009594

Copyright:© 2021 Khong et al. This is an open

access article distributed under the terms of the

Creative Commons Attribution License, which

permits unrestricted use, distribution, and

reproduction in any medium, provided the original

author and source are credited.

Data Availability Statement: All relevant data are

within the manuscript and its Supporting

Information files.

these genes. Two of the genes act as negative regulators of the panicle meristem’s identity

transition from indeterminate to determinate state while the other two act as positive reg-

ulators of this meristem fate transition. Based on the different phenotypes between over-

expressed and mutant plants, two of these genes were confirmed as playing a role in the

control of panicle architecture. These findings can be directly used to assist selection for

grain yield improvement.

Introduction

Rice yield is determined by several traits, including the number of grains carried on each pani-

cle [1,2]. This in turn is dependent on the architecture of the panicle, which consists of a series

of axes of different successive orders: rachis; primary branches; secondary branches; poten-

tially tertiary branches; and finally spikelets. Since the rice spikelet typically contains a single

fertile floret, the number of spikelets usually determines in fine the number of grains that can

be produced per panicle. In this context, genetic selection of panicle branching traits to

improve yield potential is of great importance in breeding programmes. The complexity of

panicle branching is determined by two interconnected developmental processes: the number

of axillary meristems produced along each of the panicle axes; and the rate of meristem fate

transition, which determines whether an axillary meristem keeps its indeterminacy so as to

grow into a higher-order branch (secondary or tertiary branch) or whether it becomes deter-

minate and therefore differentiates into a lateral spikelet. Various genes that control these two

processes governing panicle architecture were identified through the characterization of

mutants and QTLs, shedding light on the signalling pathways and protein interactions

involved [1–5].

Proteins with tandem peptide repeats are essential for fundamental biological functions,

through their involvement in protein complexes in which they may perform a scaffolding role

[6,7]. Among these proteins, the Ankyrin (ANK) domain containing proteins and the Tetratri-

copeptide repeat (TPR) domain containing proteins serve as multiprotein complex mediators

which function in transcriptional regulation, transport, biosynthesis of the photosynthetic sys-

tem and protein modification in plants. As a consequence, these proteins regulate diverse bio-

logical processes including disease response, abiotic stress response and other processes

important for plant growth and development [7–10]. Several genes have been characterized

that encode ANK- and TPR-containing proteins involved in reproductive development in

grasses. The rice DECREASED SPIKELET 4 gene (DES4) encoding a TPR-LRR protein ortho-

logous to BRUSHY1/TONSOKU/MGOUN3 of Arabidopsis thaliana, seems to be required for

genomic maintenance like its A. thaliana ortholog and affects reproductive meristem activity

as illustrated by the decreased spikelet number of the mutant [11–16]. The rice OsSPINDLY(OsSPY) gene, which encodes a TPR domain-containing N-acetyl glucosamine transferase,

modulates plant growth and architecture, including panicle architecture, through its activity as

a negative regulator of gibberellin (GA) signalling via DELLA proteins [17–18]. ANK-contain-

ing proteins that affect inflorescence development and meristem fate control include the tran-

scriptional co-activators ANK BTB/POZ proteins belonging to the BLADE-ON-PETIOLE

(BOP) clade [19]. The BOP gene TRU1 (for TASSELS REPLACE UPPER EARS 1) from maize

was identified as a direct target of the TEOSINTE BRANCHED1 (TB1) transcription factor

that regulates plant architecture through repression of axillary meristem outgrowth [20]. The

loss of function mutant tru1 is characterized by the presence of long branches tipped by a tas-

sel-like inflorescence in place of the short female axillary branch or ears. The TRU1 gene may

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 2 / 24

Funding: This research was sponsored by Vietnam

National Foundation for Science and Technology

Development (https://nafosted.gov.vn/en/) through

grant 106-NN.02-2016.60 to GNK, a joined funding

by Agropolis Foundation (« Investissements

d’avenir » programme (ANR-10-LABX-0001-01),

https://www.agropolis-fondation.fr) and

Fondazione Cariplo under the reference ID

EVOREPRICE 1201-004 to SJ and the CGIAR

Research program on Rice (CRP Rice, http://

ricecrp.org) to SJ. The funders had no role in study

design, data collection and analysis, decision to

publish, or preparation of the manuscript.

Competing interests: The authors have declared

that no competing interests exist.

also play a role in specification of basal leaf compartments, as observed for BOP genes from

barley, the rice crop O. sativa and the wild rice Oryza longistaminata (CUL4, OsBOP1/2/3,

OlBOP1/2/3 genes respectively) [21–23]. The CUL4 and OsBOP genes regulate tillering

and spikelet organ development [21,22], as do their orthologs in Brachipodium distachyon,

BdUNICULME4 and BdLAXATUM-A [24,25].

To resolve the individual contributions of the different morphological components to

spikelet number per panicle (SpN) in rice, we previously performed a genome-wide associa-

tion study (GWAS) on a panel of Vietnamese rice landraces leading to the identification of 29

QTLs [26]. One of these QTLs on chromosome 2, namely QTL_9, co-associated with SpN and

secondary branch number (SBN) traits. In order to validate the GWAS-derived QTL and to

identify the gene(s) of functional importance that it represents, a detailed analysis of the corre-

sponding genomic region was carried out in the present study, leading to the identification of

two genes affecting panicle architecture within a cluster encoding several different ANK and

ANK-TPR containing proteins. Promoter sequence variations, involving deletions of putative

transcription factor binding sites, correlate with the differential expression levels of the two

key ANK/ ANK-TPR genes with respect to panicle secondary branch number.

Results

Genetic validation of the GWAS-derived QTL_9

In a previous study, we identified a GWAS site on chromosome 2, designated as QTL_9, that

displayed co-localization for the characters spikelet number per panicle (SpN) and secondary

branch number per panicle (SBN) [26]. This region of 783 Kbp (positions 16571984 to

17355751) includes 77 annotated coding genes and 3 transposable elements in the O. sativa cv.

Nipponbare reference genome (Figs 1 and S1 and S2 Table). Among the 24 distinct haplotypes

in this region, two main haplotypes with contrasting phenotypic values for the two morpho-

logical traits were identified: H1 with low branching values and H2 with high branching ones

(S2 Fig). These two haplotypes were selected as they characterized a large number of accessions

and a wide genetic distance, while being mainly constituted by O. sativa ssp. indica accessions;

H1 being indica specific. For further genetic analysis, 7 and 5 indica accessions were selected

respectively from the H1 and H2 haplotype panels, which originated mainly from the Red

Fig 1. Polymorphisms between haplotypes H1 and H2 in the QTL_9 intragenic regions. Schematic view of the QTL_9 region with annotated genes. The genes which

were captured for sequencing are indicated in brown and the non-captured genes in grey. The different polymorphic sites identified by gene capture within the genes

between the haplotypes H1 and H2 are indicated by coloured dots, with a dot per polymorphic site: orange for non-synonymous coding, blue for synonymous coding,

green for frame shift, purple for codon deletion, deep red for Stop gain and light brown for Start gain.

https://doi.org/10.1371/journal.pgen.1009594.g001

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 3 / 24

River delta in the case of the H1 accessions and from the North Vietnam mountainous areas

for those of H2 (S3 Fig and S1 Table).

A gene capture experiment, based on these selected accessions from H1 and H2 haplotypes,

enabled the sequencing of 72 of the 77 genes (including upstream and downstream regions),

leading to the identification after filtering of 1035 polymorphic sites between the two haplo-

types, including both SNPs and INDELs (S2 and S3 Tables). In the captured intragenic regions,

several modifications that entailed protein changes were observed at a higher density towards

the 3’ end of QTL_9 (Fig 1 and S2 and S3 Tables).

In order to validate the GWAS site, a bi-parental population was produced by crossing the

low branching Sớm Giai Hưng Yên (G6) accession with the high branching Khẩu Nam Rinh

(G189) accession. These two accessions carry the H1 and H2 haplotypes respectively, which

are characterized by contrasting phenotypes for the SpN and SBN panicle morphological traits

(Figs 2 and S3). F2 plants (n = 275) were genotyped using CAPS markers derived from Gene

Capture-based SNP calling, leading to the selection of 49 homozygous plants over the QTL_9

region for both haplotypes. The phenotyping of the F3 plants revealed significantly divergent

values between the two haplotypes for the highly correlated SpN and SBN traits (Figs 2 and

S4A), confirming the importance of this genomic region in panicle branching. Phenotyping

for other traits revealed no significant segregation for the character primary branch number

per panicle (PBN); nor for flowering time, tiller number, or efficient tiller number (S4B and

S4C Fig).

Identification of a cluster of ANK and ANK-TPR genes differentially

expressed in QTL9

Of the 77 annotated genes in the O. sativa cv. Nipponbare reference genome in QTL_9 region,

16 were reported as expressed during early stages of panicle development according to publicly

available databases and panicle-derived RNA-seq data (Fig 3A and S2 Table) [27,28]. In order

to determine whether the difference of branching phenotype between the two haplotypes

might be associated with differential expression of genes during early panicle development,

expression profiling of the 16 genes from QTL_9 expressed in the panicle was carried out for

the 2 haplotypes. Panicle developmental staging was defined as follows: “early branching”

(from inflorescence meristem stage to panicle with elongated primary and higher order branch

development); and “late branching” (from panicle with elongated primary and secondary

branches to young flowers with differentiated organs), the first being enriched in meristems of

an indeterminate state and the second one with meristems of a determinate state. The expres-

sion of the LOC_Os02g28140 gene in panicle was not confirmed in our condition. Although

expression levels often varied between different stages of panicle development, no significant

expression differences were observed between the two haplotypes for most of the genes investi-

gated (S5 Fig). However, in a specific region of QTL_9, corresponding to a cluster of 6 genes

encoding ANK- and ANK-TPR-containing proteins and 10 additional genes, significant dif-

ferential expression was observed between the two haplotypes for 4 genes: LOC_Os02g29040,

LOC_Os02g29160, LOC_Os02g29190 and LOC_Os02g29210 (Fig 3B and 3C). Two types of

branching-related expression profile were observed: higher expression in accessions from the

low branching value H1 haplotype (LOC_Os02g29160 and LOC_Os02g29210); and higher

expression in accessions from the high branching value H2 haplotype (LOC_Os02g29040 and

LOC_Os02g29190) (Fig 3C).

Due to the differential expression between the two haplotypes of certain genes located in

the abovementioned gene cluster in the 3’ proximal region of QTL_9, we focused on this

region in subsequent investigations. The cluster is composed of 6 ANK genes and 10 additional

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 4 / 24

genes encoding several unknown expressed proteins, a protein kinase and a F-box protein, as

well as a short ANK gene (LOC_Os02g29110) considered to be non-functional (Fig 3B). Five of

the 6 ANK genes are annotated as belonging to the ANK-TPR subfamily and the last one

(LOC_Os02g29040) as belonging to the ANK-M subfamily (i.e. proteins with ANK domains

only) [8] (S6 Fig). Interestingly, the LOC_Os02g29040 gene might be derived from an ancestral

Fig 2. Panicle architecture in the F3 G6xG189 bi-parental population. (A) Box-plot with individual dots indicating spikelet number (SpN)

and secondary branch number (SBN) per panicle in the two parental accessions G6 from haplotype H1 and G189 from haplotype H2. (B) Box-

plot with individual dots indicating spikelet number (SpN) and secondary branch number (SBN) per panicle in the F3 lines from the G6xG189

bi-parental population with H1 haplotype (F3_H1) or H2 haplotype (F3_H2) in the QTL_9 region. (C) Mature panicle from G6 and G189

accessions, a F3_H1 line and a F3_H2 line. Individual dots in box-plots correspond to average values from the 3 main panicles per plant. Scale

bar: 2 cm. Statistical significance (t test p values) between the two lines or parents for the two panicle morphological traits is indicated as follows:� if p-value<0.05, �� if<0.01, ��� if<0.001.

https://doi.org/10.1371/journal.pgen.1009594.g002

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 5 / 24

ANK-TPR gene due to a mutation leading to a premature STOP codon, given that as a coding

sequence for TPR domains is present in the long 3’UTR region (S7 Fig). This STOP codon,

which is present in other japonica and indica cultivars, is also observed in the genome of the

wild relative species Oryza rufipogon (acc. W1943 from Gramene database) (S5 Table). More-

over, sequencing of the LOC_Os02g29210 cDNA showed that exon 2, as reported in the O.

sativa ssp. japonica cv. Nipponbare MSU7.0 reference genome, is intronic in the Nipponbare

and Kitaake cultivars that we used (S7 Fig).

The differential expression profiles of the ANK and ANK-TPR genes do not show any rela-

tionship that might reflect localization within the gene cluster (Fig 3B and 3C) or molecular

phylogenetic similarities (S6 Fig). Indeed, phylogenetic analysis of the ANK-TPR subfamily

genes from O. sativa (17 genes), Zea mays (4 genes) and A. thaliana (1 gene) revealed that the

rice genes clustered in QTL_9, including LOC_Os02g29040, are closely related both to each

other and to two other rice genes, LOC_Os03g42350 and LOC_Os08g13640. Thus, the gene

cluster characterised in the present study is a result of several events of tandem duplications of

an ancestral ANK-TPR gene. The closest characterised relative of the aforementioned rice

genes outside the genus Oryza is the maize gene GRMZM2G536120 (S6 Fig). The unique A.

thaliana ANK-TPR gene At3g04710/AtTPR10 (or AtTPR071 in [10]) is more closely related to

the rice gene LOC_Os05g01310 located on chromosome 5 (S6 Fig).

ANK and ANK-TPR gene cluster polymorphisms

Gene Capture data revealed a high degree of polymorphism between the two haplotypes within

and around the ANK and ANK-TPR genes (Figs S8 and 1 and S3 Table). Polymorphisms

between the two haplotypes affecting encoded protein sequences were identified in each of the

ANK and ANK-TPR genes in the cluster with the exception of LOC_Os02g29130 (S9 Fig).

Some of these polymorphic sites are non-synonymous. Of these substitutions, 4 are located in

Fig 3. Structure of the ANK and ANK-TPR gene cluster in QTL_9 and expression profiling. (A) Schematic view of the annotated genes according to the O. sativa ssp.

japonica cv. Nipponbare MSU7.0 reference genome in the 780 Kbp region corresponding to QTL_9 showing genes that are expressed (green) or not expressed (grey) in

the developing panicle according to publicly available databases and RNA-seq datasets. (B) Schematic view of the cluster of ankyrin genes in the 3’ proximal region of

QTL_9 with annotated genes shown based on the O. sativa ssp. japonica cv. Nipponbare MSU7.0 reference genome. The orientation of the genes is indicated by arrows.

(C) Histograms of expression profiling by qRT-PCR of the 6 ANK and ANK-TPR genes in two accessions from H1 haplotypes (G6 and G19) and two accessions from H2

haplotypes, based on two panicle developmental stages: “early branching” and “late branching”. Data are given as means ± SD.

https://doi.org/10.1371/journal.pgen.1009594.g003

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 6 / 24

ANK domains and 6 in TPR domains (S9 Fig). Two of the ANK domain polymorphisms

might have a functional significance as they occur at well conserved positions involved in the

determination of protein structure in LOC_Os02g29210 (T to N in H1 haplotype) and

LOC_Os02g29190 (G to E in H1 haplotype) (S9 Fig) [29,30].

All polymorphic sites identified in the coding sequences were also observed to occur within

the indica subpopulation (n = 1765) from the 3K Rice Genomes project [31], in which addi-

tional polymorphic sites could be found (S10 Fig). These missing sites were supposed to be

monomorphic between the Nipponbare reference genome and the H1 and H2 haplotypes. The

density of polymorphic sites was observed to be higher in the 5’ upstream regions of the genes

that displayed differential expression between the two haplotypes, with large deletions (from

18 to 42 bp) mainly observed in the H1 haplotype compared to the O. sativa cv. Nipponbare

reference genome (Figs 4 and S8). Polymorphisms within the ANK and ANK-TPR genes and

their promoter regions between Vietnamese haplotypes H1 and H2 were compared to data

available for the indica subpopulation of the 3K Rice Genomes project (S11 Fig). We observed

that the H1/H2 polymorphic sites were also present in the indica subpopulation, with the

exception of the large INDELs found in promoter regions in both H1 and H2 haplotypes. The

Fig 4. Promoter regions of the ANK and ANK-TPR genes with differential expression between the two haplotypes. (A) Promoter regions and the first exon of the 4

ANK and ANK-TPR genes showing differential expression between the haplotypes H1 and H2. INDELs are indicated by dotted lines: red with a deletion in H1 and green

with a deletion in H2 compared to the O. sativa ssp. japonica cv. Nipponbare MSU7.0 reference genome. The sequences and positions of polymorphic sites for the cultivar

Nipponbare (Nip) and the H1 and H2 haplotypes (H1, H2) are indicated below, with the H1 vs. Nip polymorphic sites in blue and the H2 vs. Nip polymorphic sites in

white. (B) Sequence alignment of the promoter regions showing large deletions in the two haplotypes (Indica_H1, Indica_H2) for the ANK and ANK-TPR genes from the

two O. sativa ssp. japonica Nipponbare (Japonica_Nip) and Kitaake (Japonica_Kita) genomes; the two O. sativa ssp. indica R498 and 93–11 genomes (Indica_R498,

Indica_9311 respectively); and the O. sativa aus N22 genome (Aus_N22). The positions of the putative TFBSs are indicated by red and green boxes, as well as the names of

the corresponding TFs.

https://doi.org/10.1371/journal.pgen.1009594.g004

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 7 / 24

latter were found to be generally missing or filtered out in the indica subpopulation dataset.

The H1 and H2 haplotypes are shared by approximately 16 and 18% of indica accessions

respectively and were observed to separate widely from each other within our analysis, each

clustering with specific indica haplotypes from the 3K Rice Genomes subpopulation, to which

they are therefore more closely related. Considering the full dataset from the 3K genome proj-

ect, about 90% and 80% of the accessions sharing H1 and H2 haplotypes respectively are

indica.

A detailed analysis of the promoter region of the 4 ANK and ANK-TPR genes revealed that

some haplotype-specific deletions had occurred in conserved MULE- and MITE-derived

sequences (Figs 4A and S12). Moreover, comparisons with available high-quality full genome

sequences from various rice cultivars and from the wild relative Oryza rufipogon (W1953 acces-

sion) indicated that the large deletions observed in the promoter regions of LOC_Os02g29040,

LOC_Os02g29160 and LOC_Os02g29190 in the H1 haplotype were shared with other indicaaccessions but absent from the wild species (Figs 4B and S12). In contrast, the two large dele-

tions observed in the LOC_Os02g29210 promoter were not observed in the available full rice

genome sequences and might be specific to the H2 haplotype (Figs 4B and S12). More interest-

ingly, the deletions observed in theses 4 genes occurred in predicted transcription factor bind-

ing sites (TFBSs) from different families (Figs 4B and S12). From an O. sativa ssp. japonica cv.

Nipponbare RNA-seq dataset obtained previously using a laser microdissection-based approach

[27], it was possible to identify several genes which encode TFs of the aforementioned families

and which are co-expressed with the ANK and ANK-TPR genes of interest in panicle reproduc-

tive meristems (S13 Fig).

Functional involvement of the ANK and ANK-TPR genes from QTL_9 in

panicle architecture regulation

In order to determine whether the ANK and ANK-TPR genes in QTL_9 influence panicle

architecture, over-expression and CRISPR-Cas9 genome editing approaches were used to

investigate two of these ANK and ANK-TPR genes, namely LOC_Os02g29040 and

LOC_Os02g29210, which show opposing expression patterns between the two haplotypes and

which belong to two different subfamilies. The functional analyses were carried out in O. sativassp. japonica cv. Kitaake.

Three overexpressing lines for the Nipponbare allelic form of LOC_Os02g29040 gene were

selected for phenotyping: ANK1_OX_1, _3 and _4 with median (ANK1_OX_1, ANK1_OX _4)

or high (ANK1_OX_3) expression level of the transgene in leaves and panicles (S14A Fig). In

the same way, three T1 overexpressing lines for the Nipponbare allelic form of LOC_Os02g29210gene were obtained but with low overexpression levels compared to wild-type (S14B Fig). Unfor-

tunately, it was not possible to maintain these plants in normal conditions, since they displayed

perturbed growth after transplanting and high sterility, suggesting a strong impact of the ubiqui-

tously expressed transgene on development. In contrast, the 3 overexpressing lines obtained for

LOC_Os02g29040 were characterized by higher PBN and SBN values compared to wild-type,

with the exception of the ANK1_OX_4 line which had a similar SBN but a higher SpN value

compared to wild-type (Figs 5A, S16B and S16D). These results suggest a positive role for the

LOC_Os02g29040 gene in panicle branching at both levels. The overexpression data corroborate

the observation that LOC_Os02g29040 expression is higher in haplotype H2 accessions, which

display higher panicle branching, than in haplotype H1 accessions (Fig 3C).

Using CRISPR-Cas9 technology, it was possible to generate various mutant alleles for the

two genes: ank1-1 to ank1-3 for LOC_Os02g29040; ank2-1 to ank2-3 for LOC_Os02g29210;

plus a double mutant ank1ank2-1 (S15 Fig). The ank1-1 allelic form of the LOC_Os02g29040

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 8 / 24

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 9 / 24

gene encoded a protein with a single amino acid deletion, in contrast to the ank1-2 and ank1-3mutants in which the last 4 ankyrin (ANK) domains were deleted (S15 Fig). The single amino

acid deletion in ank1-1 allele is located at a highly conserved position in the second ANK

domain of the LOC_Os02g29040 protein [29]. The ank2-1 allelic form of the LOC_Os02g29210gene encodes a protein with a deletion of 20 aa but with its ANK and TPR domains still con-

served. The ank2-2 mutant encodes a protein with a deletion of the last two ANK domains and

all the TPR domains whereas the protein encoded by the ank2-3 contains a deletion of the

third ANK domain (S15 Fig). The double mutant ank1ank2-1 specified truncated forms of the

proteins encoded both LOC_Os02g29040 and LOC_Os02g29210 with only the first two ANK

domains (S15 Fig).

The three ank1 mutants were characterised by normal architecture and growth compared

to wild-type. However, in a opposite way of the overexpressing lines, the ank1 mutants were

characterized with a lower number of secondary branches leading to a lower spikelet number,

without alterations to either primary branch number or rachis length. The three ank2 mutants

were also characterized by normal architecture and growth compared to wild-type but with a

higher number of secondary branches leading to higher spikelet number, without alterations

to either primary branch number or rachis length (Figs 5, S16B and S16D). This would suggest

a specific role for LOC_Os02g29040 and LOC_Os02g29210 in the control of secondary branch

number. Compared to ank1 and ank2 single mutants, the double mutant ank1ank2-1 is char-

acterized by higher SpN and SBN values, as seen for the ank2 single mutants, suggesting a

higher impact of LOC_Os02g29210 mutation on panicle architecture.

Overall, these data indicate that the LOC_Os02g29040 and LOC_Os02g29210 genes play a

role in the control of panicle architecture in rice.

Expression profiling of ANK and ANK-TPR genes from QTL_9 during

panicle development

As reported in [8] and in publicly available databases such as Rice-X-Pro (https://ricexpro.dna.

affrc.go.jp) or IC4R (http://www.ic4r.org), the rice ANK / ANK-TPR genes have different

expression profiles in different tissues or organs but generally display a higher expression level

in callus, in the panicle, in meristems and/or in spikelets in most cases (S17 Fig). In order to

detail their expression patterns during early panicle development (i.e. stages preceding floral

organ differentiation), a developmental time course analysis was performed on the Nipponbare

cultivar by using a single panicle per stage (Fig 6A and 6B). Genes could be categorised accord-

ing to their expression levels during two broadly defined phases of development: the indetermi-

nate phase (characterised by inflorescence and branch meristems) and the determinate phase

(characterised by spikelet and floret meristems). LOC_Os02g29040, LOC_Os02g29130 and

LOC_Os02g29140 were found to display higher expression during the determinate phase (stages

Fig 5. Involvement of LOC_Os02g29040 and LOC_Os02g29210 genes in panicle architecture variation. (A) On the left, box-plots

with individual dots of indicating spikelet number (SpN) and secondary branch number (SBN) per panicle in the different lines

overexpressing the LOC_Os02g29040 gene, in comparison with the wild-type Kitaake cultivar (WT); on the right, mature panicles

from the ANK1_OX_1 line and the wild-type. (B) On the left, box-plots with individual dots of the spikelet number (SpN) and

secondary branch number (SBN) per panicle in the different CRISPR-Cas9-derived lines for the LOC_Os02g29040 gene and the

double mutant ank1ank2 (ank1ank2_1), in comparison with a “sister” line (i.e. transgenic plant without mutation); on the right,

mature panicles from the ank1_2 line and the “sister” line (ANK1). (C) On the left, box-plots with individual dots indicating the

spikelet number (SpN) and secondary branch number (SBN) per panicle of the different CRISPR-Cas9-derived lines for the

LOC_Os02g29210 gene and the double mutant ank1ank2 (ank1ank2_1), in comparison with a “sister” line (i.e. transgenic plant

without mutation); on the right, mature panicles from the ank2_1 line and the “sister” line (ANK2). Individual dots in box-plots

correspond to average values from the 3 main panicles per plant. Scale bar: 2 cm. Statistical significance (t-test p values) between the

two lines or parents for the two panicle morphological traits is indicated as follows: NS if the test is non-significant; � if p-values

<0.05; �� if<0.01; ��� if<0.001.

https://doi.org/10.1371/journal.pgen.1009594.g005

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 10 / 24

N5 to N7) with a peak at the beginning of spikelet differentiation (stage N5). Similarly,

LOC_Os02g29160 expression was higher from the end of the indeterminate phase until floral

organ establishment (stages N4 to N7), with a peak at the spikelet differentiation stage (stage

N6). In contrast, LOC_Os02g29190 and LOC_Os02g29210 expression levels varied less between

the two main phases of the developmental time course in the Nipponbare genetic background.

TAWAWA 1 (TAW1) and LEAFY HULL STERILE 1 (LHS1) genes were used as reference

expression markers for respectively the indeterminate and determinate phases of the develop-

mental time course [32,33]. Using in situ hybridization analysis, transcripts of

LOC_Os02g29040, LOC_Os02g29160 and LOC_Os02g29210 genes were detected in the panicle

within both branch and spikelet meristems (S18 Fig).

Overall, these data show that the six ANK and ANK-TRP genes in QTL_9 in O. sativa cv.

Nipponbare are expressed during early panicle development in the different panicle meristem

types with higher levels at determinate stage (i.e. spikelet and florets) for four of them.

Fig 6. Expression pattern of ANK and ANK-TPR genes from QTL9 during panicle development. (A) Panicle staging used for the expression

profiling. N1 = rachis meristem; N2 = primary branch initiation; N3 = early primary branch elongation; N4 = primary branch elongation with

axillary meristems; N5 = spikelet meristem differentiation; N6 = floret meristem differentiation; N7 = floral organ establishment. Stages N1 to

N4 correspond to indeterminate stages, stages N5 to N7 to determinate stages. Scale: 100 μm. (B) Expression profiles of the ANK and ANK-TPRgenes during panicle development using a single panicle per stage for Fluidigm qRT-PCR. The LHS1 and TAW1 genes were used as quality

control of panicle stage sampling. Data are given as means ± SD.

https://doi.org/10.1371/journal.pgen.1009594.g006

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 11 / 24

Discussion

An ancient ANK-TPR gene cluster is contained in the QTL_9 genomic

region

Molecular characterization of QTL_9, previously identified by a GWAS study of morphologi-

cal panicle traits [26], has led us to identify a cluster of ANK- and ANK-TPR-type genes that

affect panicle architecture through modulation of secondary branch number and therefore

spikelet number per panicle and grain yield.

All six of the QTL_9 ANK genes are ANK-TPR derived, LOC_Os04g29040 being an

ANK-TPR gene with a premature STOP codon that is also observed in the wild species O. rufi-pogon. This protein subfamily shows an expanded number of members in rice compared to

other plant species, with 17 genes present in O. sativa (in the reference genome MSU7.0), 4 in

maize and tomato, and only a single locus in A. thaliana [8–10]. The expansion of the

ANK-TPR gene family in rice has occurred through internal tandem duplications as demon-

strated by ANK-TPR gene cluster organization in the rice genome [8], especially within the

ANK-TPR gene cluster in QTL_9. This suggests that an important role has been played by

these genes in the evolutionary success of rice through processes of neo- and/or sub-functiona-

lization within the expanded family, leading to enhanced adaptation and morphological diver-

sity. All of the ANK-TPR genes from O. sativa, as well as the ANK gene LOC_Os02g29040 in

QTL_9, are conserved in O. rufipogon, indicating ancient duplication events in the Oryza

genus that preceded rice domestication. The promoter sequences of these paralogues diverged

in parallel to their differential expression patterns within the rice plant as reported here and in

previous studies [8,9]. Part of this divergence relates to the differential insertion of Mutator-like transposable elements (MULEs) and Miniature Inverted-repeat Transposable Elements

(MITEs), which are DNA transposons with high copy numbers predicted to play an important

role in genome evolution and gene regulation, notably by providing de novo regulatory motifs

[34,35]. The detected MULE and MITEs in the promoters of the ANK genes from QTL_9 are

conserved in various accessions of O. sativa in both aus, japonica and indica ecotypes, as well

as in O. rufipogon, indicating their ancient origin in the genus Oryza before domestication.

Nevertheless, the allelic diversity seen in the promoter regions of the ANK and ANK-TPRgenes of QTL_9 has resulted at least in part from deletions within these ancient TEs leading to

differential loss of TF binding sites within indica accessions after domestication.

Allelic diversity of the ANK and ANK-TPR proteins

The TPR domain is a 34 amino acid long degenerate repeat corresponding to two anti-parallel

α-helices, whereas the ANK domain is a 33 amino acid long degenerated repeat with a struc-

ture of two anti-parallel α-helices followed by an anti-parallel β-sheet. Only some of the amino

acids in the ANK and TPR motifs are conserved, including hydrophobic positions necessary

for the secondary structure [29,30]. Several non-conserved amino acids are responsible for

partner molecule recognition and interaction, allowing these domains to bind diverse groups

of proteins [29,30]. In this context, it is interesting to note that certain highly conserved amino

acids of potential structural and functional importance in the ANK domains encoded by

LOC_Os02g29190 and LOC_Os02g29210 are altered in indica accessions with low branching

panicles. In the case of LOC_Os02g29210, the presence of potentially deleterious codon

changes in the allele found in low branching accessions, coupled with its higher expression

level in plants of this type, points to a possible dominant negative role for the encoded protein.

Conversely, the putatively non-functional allele of the LOC_Os02g29190 gene, which is also

found in low branching accessions, displays a low expression level, so its impact on branching

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 12 / 24

may be a result of biological inactivity and/or low expression. Finally, the phenotypic data of

the ank1-1 line carrying a single amino acid deletion in a conserved position of the second

ANK domain of the LOC_Os02g29040 protein illustrates the impact of this type of mutation

on panicle architecture. In the light of these results, we cannot rule out the possibility that pro-

tein variants specified by allelic forms of the ANK genes in QTL_9 may influence panicle

branching, although the functional effects of the allelic variants remain to be demonstrated.

The QTL_9 effect is associated with differential expression of genes

encoding ANK and ANK-TPR proteins

The importance of gene expression modulation in relation to domestication and morphologi-

cal diversification in rice is illustrated by several QTLs that affect panicle branching diversity

through differential gene expression, as a result of variations in proximal or distal regulatory

sequences or via epigenetic post-transcriptional regulation mechanisms, as seen for the cytoki-

nin oxidase/dehydrogenase OsCKX2 gene, the Squamosa Promoter Binding Protein-Like

OsSPL14 gene, the ABERRANT PANICLE ORGANIZATION 1 (APO1) gene, the gibberellin

biosynthesis enzyme GA20ox4 gene and the AP2/ERF FRIZZY PANICLE (FZP) gene [36–43].

In the same way, panicle morphological diversity associated with QTL_9 is accompanied by

variations in expression levels for 4 out of the 6 ANK and ANK-TPR genes, two of them

(LOC_Os02g29040 and LOC_Os02g29190) showing higher expression in high yield indicaaccessions while the other two (LOC_Os02g29160 and LOC_Os02g29210) display higher

expression in the low yield indica accessions. These data are corroborated by those obtained

with the overexpressing lines for LOC_Os02g29040 and the loss of function lines for

LOC_Os02g29210, which in each case produce a higher number of secondary branches and

spikelets per panicle.

Expression of the QTL_9 ANK and ANK-TPR genes was seen to vary in relation to deletions

of putative transcription factor binding sites in their proximal regulatory regions. Depending

on the target genes, these TFs might act as positive or negative regulators of their expression.

The rice ANK and ANK-TPR genes described here display co-expression during panicle devel-

opment with several TF genes from the AP2/ERF, WUSCHEL-related homeobox (WOX) and

Auxin Response Factor (ARF) families known to be involved in the regulation of inflorescence

structure [28,42,44–48]. The latter groups therefore provide good candidates for further inves-

tigations of the regulatory network in which the ANK and ANK-TPR genes operate.

QTL_9 ANK and ANK-TPR proteins may regulate meristem fate transition

during panicle development

Both the higher expression of LOC_Os02g29040 and LOC_Os02g29190 and the lower expres-

sion of LOC_Os02g29160 and LOC_Os02g29210 are associated with a higher secondary

branching in accessions of H2 indica haplotype. An inverse relationship is seen in lower sec-

ondary branched accessions of H1 indica haplotype. Consequently, the two groups of ANKand ANK-TPR genes might act respectively as activators and repressors of secondary

branching.

Panicle architecture diversity, notably in terms of spikelet number variation, can be

explained at least in part by the fine-tuning of axillary meristem fate on primary branches,

which affects the balance between secondary branch (indeterminate) and spikelet (determi-

nate) meristems. In this way, a more highly branched panicle can be obtained by a delayed or

lower rate of spikelet meristem fate acquisition [5,49]. The latter hypothesis fits with a general

model of inflorescence architecture evolution proposed on the basis of differences in the time

period required for terminal and axillary meristems to acquire floral fate [50]. In this context,

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 13 / 24

the rice ANK and ANK-TPR genes described here can be separated into two different catego-

ries in terms of their influence on meristem fate: either positive regulators of the indeterminate

to determinate fate transition leading to lower secondary branching (LOC_Os02g29160 and

LOC_Os02g29210); or negative regulators of this transition leading to higher secondary

branching (LOC_Os02g29040 and LOC_Os02g29190). According this model, the putatively

non-functional allelic form of the LOC_Os02g29210 gene may act in a dominant negative fash-

ion through binding of a biologically impaired polypeptide to partner proteins that act as nega-

tive regulators of the indeterminate to determinate meristem fate transition. On the other

hand, the fact that we were unable to obtain overexpressing lines for LOC_Os02g29210 sug-

gests that drastic, deleterious effects on plant growth and development result when ubiquitous

levels of its encoded protein are accumulated.

Since the two types of domain do not themselves have an enzymatic activity or a specific

function other than protein-protein interaction, it is likely that the ANK and ANK-TPR pro-

teins are involved in protein complexes, acting as co-chaperone or scaffold proteins. This was

recently confirmed for the sole ANK-TPR protein in A. thaliana (AtTPR10 or AtTPR071 in

[10]) acting in vitro as a molecular chaperone in high molecular weight protein complexes

with the synergistic involvement of both the ANK and TPR domains [51]. In this context, it

can be postulated that the ANK and ANK-TPR proteins in QTL_9 may act as co-chaperones

in regulatory multiprotein complexes within panicle meristems.

In conclusion, the wide diversity observed amongst the indica rice accessions in our panel,

in terms of their panicle architecture, depends in part on four genes encoding ANK- and

ANK-TPR-containing proteins that act as positive or negative regulators of the transition from

indeterminate to determinate meristem fate during inflorescence development. It will be of

great interest to elucidate relationships between the genes described here and others encoding

functionally characterized factors involved in panicle development. More specifically, future

studies should aim to identify the transcriptional regulators of the rice genes encoding ANK or

ANK-TPR proteins, as well as the molecular partners which form protein complexes with the

latter within the regulatory network controlling panicle meristem fate. Since the ANK-TPR

subfamily in rice is quite large compared to other plant species, it will be important to charac-

terize other members of this subfamily in order to better understand their functional diversity

in this species. Moreover, the QTL described in the present study will be of great interest for

yield improvement in rice breeding through modulation of panicle architecture to achieve a

higher spikelet number.

Materials and methods

Gene capture and sequence analysis

Twelve accessions were selected in a collection of O. sativa Vietnamese landraces based on

GWAS of panicle morphological traits [26]. Seven and five indica accessions were respectively

selected for the two main haplotypes with contrasting phenotypes: H1 with low panicle

branching values; and H2 with high values (S2 and S3 Figs and S1 Table). A set of 80 bp long

baits was designed by MYcroarray (www.mycroarray.com), using an overlap of 40 bp between

consecutive baits, to target 77 coding genes and upstream regions (excluding transposable ele-

ments) located between positions 16574501 and 17363439 on chromosome 2 of O. sativa using

the MSU v7 reference genome [52] (S2 Table).

DNA extraction and library construction were performed by ANDid (www.adnid.fr) using

a home-made kit. Paired end (2x250 bp) sequencing of the 12 libraries was carried out using

an Illumina Miseq machine. The initial data analysis was conducted using TOGGLe [53] for

read trimming, mapping to the O. sativa cv. Nipponbare MSU v7 reference genome and for

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 14 / 24

SNP and INDEL calling. Annotation of the variants was carried out using the SNiPlay3 web-

based application [54]. Using a Minor Allele Frequency (MAF) of 5% and SNPs/INDELs

homozygotes, without any missing data for the 12 accessions and discriminating the 2 haplo-

types, 1035 variants were investigated (S3 Table).

FASTA sequences from the Vietnamese accessions were deduced manually according to

SNP/INDEL calling and the O. sativa cv. Nipponbare reference genome. The MBKbase [55]

and Rice Genome Hub (https://rice-genome-hub.southgreen.fr) resources were used for

sequence comparisons with the 3K genome project [31] and high-quality O. sativa genomes.

Identification of putative transcription factor-binding sites (TFBSs) was carried out using

PlantPAN3.0 website facilities [56] with a cut-off of 0.8 for the “similar score” value to select

the O. sativa TFBDs. The detection of transposable element sequences was carried out using

the RiTE database website [57].

Development of a bi-parental population for genetic validation

Two indica accessions representing the 2 contrasting haplotypes (here after designed as H1

and H2) were crossed: Sớm Giai Hưng Yên G6 (low branching accession from H1 haplotype)

and Khẩu Nam Rinh G189 (high branching accession from H2 haplotype). Twenty F1 plants

were grown for checking using SSR markers (S4 Table). F2 plants (n = 275) were genotyped

using CAPS markers (S4 Table), designed based on Gene Capture-derived SNP calling in

QTL_9 region using the CAPSdetector program (https://github.com/francoissabot/

CAPSdetector). DNA extractions were carried out using the CTAB method [58]. PCR and

restriction enzyme reactions were carried out according to [59]. F3 plants (n = 49 for each hap-

lotype) were grown in lowland field conditions near Hanoi during the 2019 spring season in

0.75 m2 plots of 16 plants each. A block factor of 5 was introduced to check for possible envi-

ronmental variations within replicates, a single block containing 20 lines (i.e. 20 plots). The

three main panicles from three randomly chosen plants per plot were collected (i.e. 9 panicles/

accession/replicate). Phenotypic analysis was carried out as described in [26].

RNA sampling and expression profiling

Panicles were collected from 2 accessions each of the 2 contrasting haplotypes: G6 and G19 for

the H1 haplotype; and G189 and G205 for the H2 haplotype (S1 Table). Panicle sampling was

carried out by defining two stages: “early branching” (from inflorescence meristem stage to

panicle with elongated primary and higher order branch development); “late branching”

(from panicle with elongated primary and secondary branches to young flowers with differen-

tiated organs), the first enriched in meristems of an indeterminate state and the second in mer-

istems of a determinate state. Total RNA was extracted using a RNeasy plant mini kit (Qiagen)

from two biological replications. For the expression profiling during early panicle develop-

ment in O. sativa cv. Nipponbare, a single panicle per stage was collected for imaging (Leica

S8APO stereomicroscope in conjunction with a Leica DFC295 camera) and RNA sampling.

Total RNA was extracted using a RNeasy Micro kit (Qiagen) from three biological replications.

High-throughput qRT-PCRs using a Biomark HD Microfluidic 96×96 Dynamic Array (Flui-

digm) were carried out as described in [28]. RNA probe synthesis and in situ hybridization

were carried out as described in [60]. All the primers used for expression analysis are listed in

S4 Table.

Plasmid construction and plant transformation

LOC_Os02g29040 (ANK1) and LOC_Os02g29210 (ANK2) CDS flanked by attB sequences

were isolated from O. sativa cv. Nipponbare panicle-derived RTs using nested PCRs

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 15 / 24

(S4 Table) and cloned into the pC5300 OE vector as described previously [61]. The resulting

plasmids (PC5300.OE-ANK1 and PC5300.OE-ANK2) were used for O. sativa ssp. japonica cv.

Kitaake transformation as detailed in [62]. Single-locus and homozygous T2 lines were

selected on the basis of hygromycin resistance and qRT-PCR analysis according to [63], using

the single copy gene SPS (Sucrose Phosphate synthase) as endogenous reference gene [64].

CRISPR/Cas9 plasmid construction was carried out using the polycistronic tRNA-gRNA

(PTG-Cas9) method [65]. Two 20 nt gRNAs were designed per gene to target the ankyrin

domain-encoding exon 4 in the LOC_Os02g29040 (ANK1) and LOC_Os02g29210 (ANK2)

genes (S4 Table). A. tumefaciens bacteria carrying the pRGEB32 plasmid with PTG structures

were used for the transformation of O. sativa cv. Kitaake. CRISPR-Cas9 induced deletions

were detected by PCR (see S4 Table for primers) and sequence analysis using DSDecode and

CRISP-ID web-based tools [66,67]. T1 Cas9-free plants homozygous for the deletion were

selected for further analysis. Phenotyping was carried out on T2 plants using the three main

panicles from 6 plants per line as described in [26] and from 6 plants for the other traits (tiller

number, panicle number and flowering time).

Supporting information

S1 Fig. GWAS-derived QTL_9. Manhattan-plot and Linkage Disequilibrium (LD) heatmap

of chromosome 2 showing significant SNPs (p-value threshold 10−3) in the QTL_9 region for

the characters spikelet number (SpN) and secondary branch number (SBN) per panicle for the

two field trials performed in 2014 and 2015 according to [26]. Red and bold back lines on the

LD heat maps delimit the LD block for the GWAS peak. The significant SNPs are labelled in

blue in the LD heatmap. The lower panel shows the annotated genes according to the O. sativassp. japonica cv. Nipponbare MSU7.0 reference genome within the 783 Kbp region corre-

sponding to QTL_9, indicating genes that are expressed (green) or not expressed (grey) in the

developing panicle according the publicly available databases and RNA-seq dataset.

(PDF)

S2 Fig. Haplotype analysis of QTL_9 region in the Vietnamese landrace collection. (A)

Relationship dendrogram of the different haplotypes based on the analysis of the polymorphic

sites from QTL_9 region used for the GWAS analysis in the Vietnamese landrace collection

[26]. The proportion of indica, japonica and admixture accessions for each haplotype is indi-

cated. The two main haplotypes, H1 and H2, are indicated. (B) Box-plots of the characters

spikelet number (SpN) and secondary branch number (SBN) per panicle evaluated in 2014

and 2015 in the accessions from haplotypes H1 and H2. Statistical significance (i.e. t test p val-

ues) between the two haplotypes for the two panicle morphological traits is indicated in each

case.

(PDF)

S3 Fig. Heatmap of the panicle morphological traits in the Vietnamese landrace collection.

Euclidiean_WardD2 heatmap based on the phenotypic values obtained in 2015. Group:

genetic group of the accessions according to [68] for indica (I1 to I6), japonica (J1 to J4) and

admixture (m, Im, Jm); Zone: region from Vietnam where the accession was originating

(MRD = Mekong River Delta; SE = Southeast; CH = Central Highlands; SCC = South Cen-

tral Coast; NCC = North Central Coast; RRD = Red River Delta; NW = Northwest; NE =

Northeast; u = unknown); Trait types: red for number related traits (PBN, primary branch

number; SBN, secondary branch number; SpN; spikelet number), white for length-related

traits (RL, rachis length; PBL, primary branch average length; PBintL, average primary branch

internode length; SBL, secondary branch average length; SBintL, average secondary branch

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 16 / 24

internode length). The accessions from haplotypes H1 and H2 are highlighted in blue and

green respectively.

(PDF)

S4 Fig. Characterization of the bi-parental G6xG189 population. (A) Panicle morphological

trait analysis in F3 plants homozygous for the two haplotypes in QTL_9 region. Left panel:

Principal Component Analysis (PCA) analysis for the two first axis. Right upper panel: correla-

tion plots of the different panicle morphological traits. Right lower panel: correlation between

the SpN and the SBN traits in the two F3 subpopulations (i.e. for haplotype H1 and H2). (B)

Box-plots of the panicle morphological trait values in the F3 lines from the G6xG189 bi-paren-

tal population with H1 haplotype (F3_H1) or H2 haplotype (F3_H2) in the QTL_9 region:

rachis length (RL), average secondary branch length (SBL), average primary branch length

(PBL), primary branch number (PBN), secondary branch internode average length (SBintL),

average primary branch internode length (PBintL). (C) Box-plots for tiller number (TN), effi-

cient tiller number (eTN), flowering date of the first panicle (FTF), flowering date for 50% of

the plants (FT_50). Statistical significance (i.e. t test p values) between the two haplotypes for

the two panicle morphological traits is indicated.

(PDF)

S5 Fig. Expression profiling of the panicle expressed non-ANK or ANK-TPR genes in

QTL_9. (A) schematic view of the 780 Kbp QTL_9 region showing annotated genes according

to the O. sativa ssp. japonica cv. Nipponbare MSU7.0 reference genome. Genes expressed in

the developing panicle are indicated in green. The ANK and ANK-TPR genes are indicated by

red asterisks. Histograms illustrate the expression profiling data by qRT-PCR of the panicle-

expressed genes which do not belong to ANK or ANK-TPR families in two accessions from H1

haplotypes (G6 and G19) and two accessions from H2 haplotypes (G189 and G205). Two pani-

cle developmental stages were considered: “early branching” and “late branching”. (B) Quality

control of the panicle sampling. Histograms of expression profiling by qRT-PCR analysis of

LHS1 and TAW1 genes known to be expressed in late and early stages of panicle development

respectively.

(PDF)

S6 Fig. Ankyrin gene cluster organization, protein structure and phylogenetic relationship.

Phylogenetic tree using the Maximum Likelihood method and JTT matrix-based model in

conjunction with amino acid sequence alignment of the ANK domain of the ANK-TPR pro-

teins from O. sativa (green dots), Z. mays (yellow dots) and A. thaliana (blue dot). Bootstrap

values (1000 tests) are shown next to the branches. Alignment and phylogenetic tree were car-

ried out using MEGA X software [69]. A schematic view of the organization of the ANK gene

cluster in QTL_9 based on the O. sativa MSU7.0 reference genome is shown on the right of the

tree, the grey lines connecting the corresponding genes between the cladogram and the cluster.

The orientation of genes is indicated by arrows. The structure of the 6 ANK and ANK-TPR

proteins from the cluster is indicated.

(PDF)

S7 Fig. Annotation of LOC_Os02g29040 and LOC_Os02g29210 genes. The structure of

the two genes is indicated as defined in MSU7.0 O. sativa reference genome. The sequence

of the 3’UTR region of LOC_Os02g29040 is indicated to illustrate the presence of a TPR

domain coding sequence downstream the STOP codon. The sequence of the 5’ part of the

LOC_Os02g29210 gene is indicated to illustrate the alternative annotation (i.e. without second

exon from MSU7.0 reference genome) of this gene in the Nipponbare accession used in

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 17 / 24

the lab.

(PDF)

S8 Fig. Genomic structure and polymorphism of the ANK and ANK-TPR genes in QTL_9.

Schematic view of the exon-intron structure of the ANK genes. Blue box: coding exon; white

box: non-coding exon; blue line: intron; black line: promoter region. The structure of the

LOC_Os02g29210 gene was corrected according the in lab sequence. Polymorphic sites are

indicated with a black vertical line for SNPs and triangles for INDELs. INDEL lengths are indi-

cated in bp. SNPs causing non-synonymous mutations are indicated in red and corresponding

variations at nucleic and amino acid levels are indicated.

(PDF)

S9 Fig. Alignment of the deduced amino acid sequences from the ANK and ANK-TPRgenes coding sequences. Alignment of the sequences from the O. sativa ssp. japonica cv. Nip-

ponbare MSU7.0 reference genome and deduced sequences from the Vietnamese indica acces-

sions from the H1 and H2 haplotypes. The ANK and TPR domains are highlighted in green

and yellow respectively. A space was introduced between the consecutive TPR domains in the

alignment. Modified amino acid in H1 or H2 accessions compared to Nipponbare are

highlighted in light green. The positions of non-synonymous polymorphic sites are indicated

in grey for amino acid changes within the same aliphatic group and in orange for amino acid

changes between different aliphatic groups. The non-synonymous substitutions at highly con-

served sites in ANK domain are boxed in red. The region missing from the LOC_Os02g29210

encoded protein sequence deduced from the in lab Nipponbare accession (compared with the

published Nipponbare reference) is indicated in italics.

(PDF)

S10 Fig. Polymorphism in the coding sequence of the ANK and ANK-TPR genes between

the two haplotypes in comparison with the indica subpopulation from the 3K genome

project. Positions of polymorphic genes are indicated on the schematic view of the gene struc-

ture. The sequence and position of polymorphic sites for the cultivar Nipponbare (Nip) and

the H1 and H2 haplotypes (H1_Lv, H2_Hv) are indicated below in comparison with the differ-

ent haplotypes (GID) from the indica subpopulation according the data available in the

MBKbase website facilities (http://www.mbkbase.org/rice). The number of cultivars for each

haplotype is indicated on the right of the table.

(PDF)

S11 Fig. Promoter regions of the ANK and ANK-TPR genes with differential expression

between the two haplotypes in comparison with the indica subpopulation from the 3K

genome project. On the left, promoter regions and the first exon (E1) of the 4 ANK and

ANK-TPR genes showing differential expression between the haplotypes H1 and H2. SNPs are

indicated by vertical black lines. INDELs are indicated by dotted lines: red for a deletion in H1

and green for a deletion in H2 compared to the O. sativa ssp. japonica cv. Nipponbare MSU7.0

reference genome. The sequence and positions of polymorphic sites for the cultivar Nippon-

bare (Nip) and the H1 and H2 haplotypes (H1_Lv, H2_Hv) are indicated below in comparison

with the different haplotypes (GID) from the indica subpopulation according the data available

in the MBKbase website facilities (http://www.mbkbase.org/rice). The number of cultivars for

each haplotype is indicated on the right of the table. On the right, relationship tree between the

different haplotypes using common SNPs in the promoter regions of the ANK and ANK-TPRgenes in conjunction with the Neighbor-joining method.

(PDF)

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 18 / 24

S12 Fig. Alignment of the promoter regions of ANK and ANK-TPR genes displaying differ-

ential expression between the two haplotypes. Sequence alignment of the promoter regions

of the two haplotypes (H1_Lv, H2_Hv) of the ANK genes with the high-quality full genome

sequences from Oryza rufipogon (W1953 accession) and rice cultivars: the two O. sativa ssp.

japonica varieties Nipponbare (Nip) and Kitaake (Kita), the two O. sativa ssp. indica varieties

R498 and 93–11 and the O. sativa aus variety N22. The positions of putative TFBSs are indi-

cated by blue boxes, along with the names of the corresponding TFs. The positions of MULE

and MITES-derived regions in the promoter regions are indicated by colour-filled boxes. Poly-

morphic sites between H1 and H2 accessions are indicated in orange. Polymorphic sites which

are monomorphic for H1 and H2 are indicated in light yellow.

(DOCX)

S13 Fig. Transcription factor-encoding genes that are co-expressed with ANK and

ANK-TPR genes. Expression profiles of LOC_Os02g29040 (A), LOC_Os02g29160 (B),

LOC_Os02g19190 (C) and LOC_Os2g29210 (D) genes and the co-expressed transcription fac-

tor-encoding genes in the laser-dissected panicle meristems according to [27]. M1 = rachis

meristem; M2 = primary branch meristem; M3 = elongated primary branch meristem with

axillary meristems; M4 = spikelet meristem. The list of TF genes was defined according the list

of genes associated with the putative TBDSs (with a similar score� 0.8) based on the search in

PlantPAN3.0 website facilities.

(PDF)

S14 Fig. Expression profiling of the overexpressing transgenic lines for LOC_Os02g29040and LOC_Os02g29210. Expression profiling of LOC_Os02g29040 (A) and LOC_Os02g29210(B) in the corresponding overexpressing transgenic lines in comparison with wild-type Kitaake

cultivar in leaf and panicle tissues as indicated. Statistical significance (t test p-values) between

the lines and the wild-type is indicated as follows: ns if the test is non-significant, �� if p-values

<0.01, ��� if <0.001.

(PDF)

S15 Fig. CRISPR-Cas9-induced deletions in LOC_Os02g29040 and LOC_Os02g29210 in

the single and double gene targeted lines. (A) Alignment of the nucleic sequence of wild-type

Kitaake with the different CRISPR-Cas9-induced alleles. (B) Alignment of the deduced amino

acid sequence of the wild-type Kitaake, with those of the different CRISPR-Cas9-induced

alleles. (C) Protein structure of the wild-type Kitaake form and the different CRISPR-Cas9-in-

duced alleles.

(PDF)

S16 Fig. Phenotypic characterization of the CRISPR-cas9 transgenic lines. (A) Plant archi-

tecture at reproductive stages for an ANK1_OE line (i.e. LOC_Os02g29040 overexpressing

line), an ank1 line and an ank2 line (CRISPR-Cas9-derived lines for the LOC_Os02g29040 and

LOC_Os02g29210 genes respectively) in comparison with Kitaake wild-type. (B) Box-plots

with individual data points showing primary branch number (PBN) per panicle in the differ-

ent ANK1_OE lines, in the ank1 and ank2 single mutant lines and in the double mutant

(ank1ank2_1), in comparison with a “sister” line or the wild-type Kitaake cultivar (WT). (C)

Box-plots with individual data points showing rachis length (RL) per panicle in the different

ANK1_OE lines, in the ank1 and ank2 single mutant lines and in the double mutant

(ank1ank2_1), in comparison with a “sister” line or the wild-type Kitaake cultivar (WT). Statis-

tical significance (i.e. t test p values) between the two lines or parents for the two panicle mor-

phological traits is indicated as follows: NS if the test is non-significant, � if p-values <0.05, ��

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 19 / 24

if<0.01, ��� if <0.001.

(PDF)

S17 Fig. Expression profiling of the ANK and ANK-TPR genes in O. sativa cv. Nipponbare.

Expression profiles of the ANK genes in various tissues or organs according to data available in

IC4R website (http://www.ic4r.org). TPM: transcripts per million.

(PDF)

S18 Fig. Spatial localization of LOC_Os02g29040, LOC_Os02g29160 and LOC_Os02g29210transcripts in the young developing panicle using in situ hybridization. The histone H4

transcripts (H4) were used as a positive control of the specificity of in situ hybridization condi-

tions. SpM: spikelet meristem; BM: branch meristem. Scale: 100 μm.

(PDF)

S1 Table. List of the Vietnamese landrace accessions used for the Gene Capture analysis.

Zone: Vietnamese region where the accession was originating (MRD = Mekong River Delta;

SE = Southeast; CH = Central Highlands; SCC = South Central Coast; NCC = North Cen-

tral Coast; RRD = Red River Delta; NW = Northwest; NE = Northeast; u = unknown);

Group: genetic group of the accessions according to [68] for indica (I1 to I6) and admixture

(m, Im, Jm); Ecosystem: ecosystem from where the accession originates (RL = rainfed lowland;

UP = upland; u = unknown).

(DOCX)

S2 Table. List of annotated and captured genes in QTL_9. Relevant information relating to

the detection of polymorphic sites within genes and their promoter regions and their impact

on protein sequences. Also shown is information on the expression of these genes during early

stages of panicle development, based on publicly available data set and local RNA seq dataset.

Applicable categories are indicated by green cells.

(DOCX)

S3 Table. List of filtered SNPs/INDELs from captured genes. Filtered SNPs and INDELs

with a MAF of 5%, homozygotes and polymorphs between the two haplotypes (considering all

the tested accessions for each haplotype).

(XLSX)

S4 Table. List of the primers used in this study.

(DOCX)

S5 Table. List of ANK-TPR genes from O. sativa, A. thaliana and Z. mays. The names of

ANK-TPR genes according to [10] are indicated, as well as the features of the gene and its

encoded protein (number of introns, protein, protein length, protein molecular weight, PI).

The predicted sub-cellular localization is indicated according to [10]. O. sativa genes men-

tioned in [10] but not maintained in the MSU7.0 version of the O. sativa genome are indicated

in grey. The presence of O. rufipogon orthologs is indicated for the corresponding O. sativagenes. O. sativa genes from QTL_9 are highlighted in orange.

(DOCX)

Acknowledgments

The authors would like to thank Laurence Albar (IRD, France) for support with bi-parental

population genotyping, Francois Sabot (IRD, France) for the use of the CAPSdetector pro-

gram, Brigitte Courtois (CIRAD, France) for providing SSR markers for indica genotyping

and Celine Cardi and Ronan Rivallan (Grand plateau technique regional de genotypage,

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 20 / 24

CIRAD, Montpellier) for support with high-throughput qRT-PCR. We also thank James Tre-

gear (IRD, France) for his careful reading and language editing of the manuscript.

Author Contributions

Conceptualization: Giang Ngan Khong, Stefan Jouannic.

Data curation: Giang Ngan Khong, Stefan Jouannic.

Formal analysis: Nhu Thi Le, Helene Adam, Stefan Jouannic.

Funding acquisition: Giang Ngan Khong, Stefan Jouannic.

Investigation: Nhu Thi Le, Mai Thi Pham, Helene Adam, Carole Gauron, Hoa Quang Le,

Dung Tien Pham, Kelly Colonges.

Methodology: Stefan Jouannic.

Project administration: Giang Ngan Khong.

Resources: Xuan Hoi Pham, Michel Lebrun.

Supervision: Xuan Hoi Pham, Vinh Nang Do, Michel Lebrun.

Visualization: Stefan Jouannic.

Writing – original draft: Stefan Jouannic.

Writing – review & editing: Giang Ngan Khong, Stefan Jouannic.

References1. Xing Y, Zhang Q. Genetic and molecular bases of rice yield. Ann Rev Plant Biol. 2010; 61: 421–42.

https://doi.org/10.1146/annurev-arplant-042809-112209 PMID: 20192739

2. Ikeda M, Miura K, Aya K, Kitano H, Matsuoka M. Genes offering the potential for designing yield-related

traits in rice. Curr Opin Plant Biol. 2013; 16: 213–220. https://doi.org/10.1016/j.pbi.2013.02.002 PMID:

23466256

3. Wang Y, Li J. Branching in rice. Curr Opin Plant Biol. 2011; 14: 94–99. https://doi.org/10.1016/j.pbi.

2010.11.002 PMID: 21144796

4. Bai X, Wu B, Xing Y. Yield-related QTLs and their applications in rice genetic improvement. J Integr

Plant Biol 2012; 54: 300–311. https://doi.org/10.1111/j.1744-7909.2012.01117.x PMID: 22463712

5. Kyozuka J, Tokunaga H, Yoshida A. Control of grass inflorescence form by the fine-tuning of meristem

phase change. Curr Opin Plant Biol. 2014; 17: 110–115. https://doi.org/10.1016/j.pbi.2013.11.010

PMID: 24507502

6. Zeytuni N, Zarivach R. Structural and Functional Discussion of the Tetra-Trico-Peptide Repeat, a Pro-

tein Interaction Module. Structure. 2012; 20: 397–405. https://doi.org/10.1016/j.str.2012.01.006 PMID:

22404999

7. Sharma M, Pandey GK. Expansion and Function of Repeat Domain Proteins During Stress and Devel-

opment in Plants. Front Plant Sci. 2016; 6: 609–615. https://doi.org/10.3389/fpls.2015.01218 PMID:

26793205

8. Huang J, Zhao X, Yu H, Ouyang Y, Wang L, Zhang Q. 2009. The ankyrin repeat gene family in rice:

genome-wide identification, classification and expression profiling. Plant Mol Biol. 2009; 71: 207–226.

https://doi.org/10.1007/s11103-009-9518-6 PMID: 19609685

9. Vo KTX, Kim CY, Chandran AKN, Jung KH, An G, Jeon J-S. Molecular insights into the function of

ankyrin proteins in plants. J Plant Biol. 2015; 58: 271–284.

10. Wei K, Han P. Comparative functional genomics of the TPR gene family in Arabidopsis, rice and maize.

Mol Breed. 2017; 37: 152.

11. Guyomarc’h S, Vernoux T, Traas J, Zhou DX, Delarue M. MGOUN3, an Arabidopsis gene with Tetratri-

coPeptide-Repeat-related motifs, regulates meristem cellular organization. J Exp Bot. 2004; 55: 673–

684. https://doi.org/10.1093/jxb/erh069 PMID: 14966212

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 21 / 24

12. Takeda S, Tadele Z, Hofmann I, Probst AV, Angelis KJ, Kaya H, et al. BRU1, a novel link between

responses to DNA damage and epigenetic gene silencing in Arabidopsis. Genes Dev. 2004; 18: 782–

793. https://doi.org/10.1101/gad.295404 PMID: 15082530

13. Suzuki T, Nakajima S, Inagaki S, Hirano-Nakakita M, Matsuoka K, Demura T, et al. TONSOKU is

expressed in S phase of the cell cycle and its defect delays cell cycle progression in Arabidopsis. Plant

Cell Physiol. 2005; 46: 736–742. https://doi.org/10.1093/pcp/pci082 PMID: 15746155

14. Guyomarc’h S, Benhamed M, Lemonnier G, Renou JP, Zhou DX, Delarue M. MGOUN3: evidence for

chromatin-mediated regulation of FLC expression. J Exp Bot. 2006; 57: 2111–2119. https://doi.org/10.

1093/jxb/erj169 PMID: 16728410

15. Batzenschlager M, Schmit AC, Herzog E, Fuchs J, Schubert V, Houlne G, et al. MGO3 and GIP1 act

synergistically for the maintenance of centromeric cohesion. Nucleus. 2017; 8: 98–105. https://doi.org/

10.1080/19491034.2016.1276142 PMID: 28033038

16. Ni S, Li Z, Ying J, Zhang J, Chen H. Decreased Spikelets 4 Encoding a Novel Tetratricopeptide Repeat

Domain-Containing Protein Is Involved in DNA Repair and Spikelet Number Determination in Rice.

Genes. 2019; 10: 214–219. https://doi.org/10.3390/genes10030214 PMID: 30871267

17. Shimada A, Ueguchi-Tanaka M, Sakamoto T, Fujioka S, Takatsuto S, Yoshida S, et al. The rice SPIN-

DLY gene functions as a negative regulator of gibberellin signaling by controlling the suppressive func-

tion of the DELLA protein, SLR1, and modulating brassinosteroid synthesis. Plant J. 2006; 48: 390–

402. https://doi.org/10.1111/j.1365-313X.2006.02875.x PMID: 17052323

18. Yano K, Morinaka Y, Wang F, Huang P, Takehara S, Hirai T, et al. GWAS with principal component

analysis identifies a gene comprehensively controlling rice architecture. Proc Natl Acad Sci, USA. 2019;

116: 21262–21267. https://doi.org/10.1073/pnas.1904964116 PMID: 31570620

19. Khan M, Xu H, Hepworth SR. BLADE-ON-PETIOLE genes: setting boundaries in development and

defense. Plant Sci. 2014;215–216: 157–171. https://doi.org/10.1016/j.plantsci.2014.09.010 PMID:

25443848

20. Dong Z, Li W, Unger-Wallace E, Yang J, Vollbrecht E, Chuck G. Ideal crop plant architecture is medi-

ated by tassels replace upper ears1, a BTB/POZ ankyrin repeat gene directly targeted by TEOSINTE

BRANCHED1. Proc Natl Acad Sci, USA. 2017; 114: E8656–E8664. https://doi.org/10.1073/pnas.

1714960114 PMID: 28973898

21. Tavakol E, Okagaki R, Verderio G, Shariati J V, Hussien A, Bilgic H, et al. The barley Uniculme4 gene

encodes a BLADE-ON-PETIOLE-like protein that controls tillering and leaf patterning. Plant Physiol.

2015; 168: 164–174. https://doi.org/10.1104/pp.114.252882 PMID: 25818702

22. Toriba T, Tokunaga H, Shiga T, Nie F, Naramoto S, Honda E, et al. BLADE-ON-PETIOLE genes tem-

porally and developmentally regulate the sheath to blade ratio of rice leaves. Nature Comm. 2019; 10:

619. https://doi.org/10.1038/s41467-019-08479-5 PMID: 30728357

23. Toriba T, Tokunaga H, Nagasawa K, Nie F, Yoshida A, Kyozuka J. Suppression of Leaf Blade Develop-

ment by BLADE-ON-PETIOLE Orthologs Is a Common Strategy for Underground Rhizome Growth.

Curr Biol. 2020; 30: 509–516.e3. https://doi.org/10.1016/j.cub.2019.11.055 PMID: 31956025

24. Jost M, Taketa S, Mascher M, Himmelbach A, Yuo T, Shahinnia F, et al. A Homolog of Blade-On-Peti-

ole 1 and 2 (BOP1/2) Controls Internode Length and Homeotic Changes of the Barley Inflorescence.

Plant Physiol. 2016; 171: 1113–1127. https://doi.org/10.1104/pp.16.00124 PMID: 27208226

25. Magne K, Liu S, Massot S, Dalmais M, Morin H, Sibout R, et al. Roles of BdUNICULME4 and BdLAXA-

TUM-A in the non-domesticated grass Brachypodium distachyon. Plant J. 2020; 103: 645–659. https://

doi.org/10.1111/tpj.14758 PMID: 32343459

26. Ta KN, Khong NG, Ha TL, Nguyen DT, Mai DC, Hoang TG, et al. A genome-wide association study

using a Vietnamese landrace panel of rice (Oryza sativa) reveals new QTLs controlling panicle morpho-

logical traits. BMC Plant Biol. 2018; 18: 282. https://doi.org/10.1186/s12870-018-1504-1 PMID:

30428844

27. Harrop TWR, Ud Din I, Gregis V, Osnato M, Jouannic S, Adam H, et al. Gene expression profiling of

reproductive meristem types in early rice inflorescences by laser microdissection. Plant J. 2016; 86: 75–

88. https://doi.org/10.1111/tpj.13147 PMID: 26932536

28. Harrop TWR, Mantegazza O, Luong AM, Bethune K, Lorieux M, Jouannic S, et al. A set of AP2-like

genes is associated with inflorescence branching and architecture in domesticated rice. J Exp Bot.

2019; 70: 5617–5629. https://doi.org/10.1093/jxb/erz340 PMID: 31346594

29. Mosavi LK, Cammett TJ, Desrosiers DC, Peng ZY. The ankyrin repeat as molecular architecture for

protein recognition. Prot Sci. 2004; 13: 1435–1448. https://doi.org/10.1110/ps.03554604 PMID:

15152081

30. Magliery TJ, Regan L. Beyond consensus: statistical free energies reveal hidden interactions in the

design of a TPR motif. J Mol Biol. 2004; 343: 731–745. https://doi.org/10.1016/j.jmb.2004.08.026

PMID: 15465058

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 22 / 24

31. Wang W, Mauleon R, Hu Z, et al. Genomic variation in 3,010 diverse accessions of Asian cultivated

rice. Nature. 2018; 557: 43–49. https://doi.org/10.1038/s41586-018-0063-9 PMID: 29695866

32. Kobayashi K, Maekawa M, Miyao A, Hirochika H, Kyozuka J. PANICLE PHYTOMER2 (PAP2), encod-

ing a SEPALLATA subfamily MADS-Box protein, positively controls spikelet meristem identity in rice.

Plant Cell Physiol. 2010; 51: 47–57. https://doi.org/10.1093/pcp/pcp166 PMID: 19933267

33. Yoshida A, Sasao M, Yasuno N, Takagi K, Daimon Y, Chen R, et al. TAWAWA1, a regulator of rice inflo-

rescence architecture, functions through the suppression of meristem phase transition. Proc Natl Acad

Sci, USA. 2013; 110: 767–772. https://doi.org/10.1073/pnas.1216151110 PMID: 23267064

34. Jiang N, Bao Z, Zhang X, Eddy SR, Wessler SR. Pack-MULE transposable elements mediate gene

evolution in plants. Nature. 2004; 431: 569–573. https://doi.org/10.1038/nature02953 PMID: 15457261

35. Lu C, Chen J, Zhang Y, Hu Q, Su W, Kuang H. Miniature Inverted-Repeat Transposable Elements

(MITEs) Have Been Accumulated through Amplification Bursts and Play Important Roles in Gene

Expression and Species Diversity in Oryza sativa. Mol Biol Evol. 2012; 29: 1005–1017. https://doi.org/

10.1093/molbev/msr282 PMID: 22096216

36. Ashikari M, Sakakibara H, Lin S, Yamamoto T, Takashi T, Nishimura A, et al. Cytokinin oxidase regu-

lates rice grain production. Science. 2005; 309: 741–745. https://doi.org/10.1126/science.1113373

PMID: 15976269

37. Jiao Y, Wang Y, Xue D, Wang J, Yan M, Liu G, et al. Regulation of OsSPL14 by OsmiR156 defines

ideal plant architecture in rice. Nature Genet. 2010; 42: 541–544. https://doi.org/10.1038/ng.591 PMID:

20495565

38. Miura K, Ikeda M, Matsubara A, Song XJ, Ito M, Asano K, et al. OsSPL14 promotes panicle branching

and higher grain productivity in rice. Nature Genet. 2010; 42: 545–549. https://doi.org/10.1038/ng.592

PMID: 20495564

39. Wang J, Xu H, Li N, Fan F, Wang L, Zhu Y, Li S. Artificial Selection of Gn1a Plays an Important role in

Improving Rice Yields Across Different Ecological Regions. Rice. 2015; 8: 37. https://doi.org/10.1186/

s12284-015-0071-4 PMID: 26677125

40. Bai X, Huang Y, Hu Y, Liu H, Zhang B, Smaczniak C, et al. Duplication of an upstream silencer of FZP

increases grain yield in rice. Nature Plants. 2017; 3: 885–893. https://doi.org/10.1038/s41477-017-

0042-4 PMID: 29085070

41. Fujishiro Y, Agata A, Ota S, Ishihara R, Takeda Y, Kunishima T, et al. Comprehensive panicle pheno-

typing reveals that qSrn7/FZP influences higher-order branching. Scientific Rep. 2018; 8: 12511.

https://doi.org/10.1038/s41598-018-30395-9 PMID: 30131566

42. Huang Y, Zhao S, Fu Y, Sun H, Ma X, Tan L, et al. Variation in the regulatory region of FZP causes

increases in secondary inflorescence branching and grain yield in rice domestication. Plant J. 2018; 96:

716–733. https://doi.org/10.1111/tpj.14062 PMID: 30101570

43. Agata A, Ando K, Ota S, Kojima M, Takebayashi Y, Takehara S, et al. Diverse panicle architecture

results from various combinations of Prl5/GA20ox4 and Pbl6/APO1 alleles. Comm Biol. 2020; 3: 302.

https://doi.org/10.1038/s42003-020-1036-8 PMID: 32528064

44. Yan WH, Wang P, Chen HX, Zhou HJ, Li QP, Wang CR, et al. A major QTL, Ghd8, plays pleiotropic

roles in regulating grain productivity, plant height, and heading date in rice. Mol Plant. 2011; 4: 319–330.

https://doi.org/10.1093/mp/ssq070 PMID: 21148627

45. Lee D-Y, An G. Two AP2 family genes, SUPERNUMERARY BRACT (SNB) and OsINDETERMINATE

SPIKELET 1 (OsIDS1), synergistically control inflorescence architecture and floral meristem establish-

ment in rice. Plant J. 2011; 69: 445–461. https://doi.org/10.1111/j.1365-313X.2011.04804.x PMID:

22003982

46. Wang W, Li G, Zhao J, Chu H, Lin W, Zhang D, et al. Dwarf Tiller1, a Wuschel-related

homeobox transcription factor, is required for tiller growth in rice. PLoS Genet. 2014; 10: e1004154.

https://doi.org/10.1371/journal.pgen.1004154 PMID: 24625559

47. Bai X, Huang Y, Mao D, Wen M, Zhang L, Xing Y. Regulatory role of FZP in the determination of panicle

branching and spikelet formation in rice. Scientific Rep. 2015; 6: 19022.

48. Fang F, Ye S, Tang J, Bennett MJ, Liang W. DWT1/DWL2 act together with OsPIP5K1 to regulate plant

uniform growth in rice. New Phytol. 2020; 225: 1234–1246. https://doi.org/10.1111/nph.16216 PMID:

31550392

49. Ta KN, Adam H, Staedler YM, Schonenberger J, Harrop T, Tregear J, et al. Differences in meristem

size and expression of branching genes are associated with variation in panicle phenotype in wild and

domesticated African rice. EvoDevo. 2017; 8: 394.

50. Prusinkiewicz P, Erasmus Y, Lane B, Harder LD, Coen E. Evolution and development of inflorescence

architectures. Science. 2007; 316: 1452–1456. https://doi.org/10.1126/science.1140429 PMID:

17525303

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 23 / 24

51. Paeng SK, Kang CH, Chi YH, Chae HB, Lee ES, Park JH, et al. AtTPR10 Containing Multiple ANK and

TPR Domains Exhibits Chaperone Activity and Heat-Shock Dependent Structural Switching. Applied

Sci. 2020; 10: 1265–1214.

52. Ouyang S, Zhu W, Hamilton J, Lin H, Campbell M, Childs K, et al. 2007. The TIGR rice genome annota-

tion resource: improvements and new features. Nucl Acids Res. 2007; 35: D883–D887. https://doi.org/

10.1093/nar/gkl976 PMID: 17145706

53. Tranchant-Dubreuil C, Ravel S, Monat C, Sarah G, Diallo A, Helou L, et al. TOGGLe, a flexible frame-

work for easily building complex workflows and performing robust large-scale NGS analyses. bioRxiv

[preprint]. 2018 bioRxiv 245480; Available from: https://www.biorxiv.org/content/10.1101/245480.

54. Dereeper A, Homa F, Andres G, Sempere G, Sarah G, Hueber Y, et al. SNiPlay3: a web-based applica-

tion for exploration and large scale analyses of genomic variations. Nucl Acids Res. 2015; 43: W295–

300. https://doi.org/10.1093/nar/gkv351 PMID: 26040700

55. Peng H, Wang K, Chen Z, Cao Y, Gao Q, Li Y, et al. MBKbase for rice: an integrated omics knowledge-

base for molecular breeding in rice. Nucl Acids Res. 2020; 48: D1085–D1092. https://doi.org/10.1093/

nar/gkz921 PMID: 31624841

56. Chow CN, Lee TY, Hung YC, Li GZ, Tseng KC, Liu YH, et al. PlantPAN3.0: a new and updated resource

for reconstructing transcriptional regulatory networks from ChIP-seq experiments in plants. Nucl Acids

Res. 2019; 47: D1155–D1163. https://doi.org/10.1093/nar/gky1081 PMID: 30395277

57. Copetti D, Zhang J, El Baidouri M, et al. 2015. RiTE database: a resource database for genus-wide rice

genomics and evolutionary biology. BMC Genomics. 2015; 16: 538. https://doi.org/10.1186/s12864-

015-1762-3 PMID: 26194356

58. Doyle JJ, Doyle JL. Isolation of plant DNA from fresh tissue. Focus. 1990; 12: 13–15.

59. Baumbusch LO, Sundal IK, Hughes DW, Galau GA, Jakobsen KS. Efficient protocols for CAPS-based

mapping in Arabidopsis. Plant Mol Biol Rep. 2001; 19: 137–149.

60. Adam H, Marguerettaz M, Qadri R, Adroher B, Richaud F, Collin M, et al. 2011. Divergent Expression

Patterns of miR164 and CUP-SHAPED COTYLEDON Genes in Palms and Other Monocots: Implica-

tion for the Evolution of Meristem Function in Angiosperms. Mol Biol Evol. 2011; 28: 1439–1454. https://

doi.org/10.1093/molbev/msq328 PMID: 21135149

61. Schmidt R, Schippers JH, Welker A, Mieulet D, Guiderdoni E, et al. Transcription factor OsHsfC1b regu-

lates salt tolerance and development in Oryza sativa ssp. japonica. AoB PLANTS. 2012; pls011. https://

doi.org/10.1093/aobpla/pls011 PMID: 22616023

62. Sallaud C, Meynard D, van Boxtel J, Gay C, Bès M, Brizard JP, et al. Highly efficient production and

characterization of T-DNA plants for rice (Oryza sativa L.) functional genomics. Theor Appl Genet.

2003; 106: 1396–408. https://doi.org/10.1007/s00122-002-1184-x PMID: 12677401

63. Zhang Y, Liu Y, Zhang J, Wang G, Wang J, Liu Y. Assessment of transgene copy number and zygosity

of transgenic maize overexpressing Cry1Ie gene with SYBR Green qRT-PCR. In Vitro Cell Dev Biol—

Plant. 2015; 51: 125–134.

64. Ding J, Jia J, Yang L, Wen H, Zhang C, Liu W, et al. Validation of a rice specific gene, sucrose phos-

phate synthase, used as the endogenous reference gene for qualitative and real-time quantitative PCR

detection of transgenes. J Agri Food Chem. 2004; 52: 3372–3377. https://doi.org/10.1021/jf049915d

PMID: 15161200

65. Xie K, Minkenberg B, Yang Y. Boosting CRISPR/Cas9 multiplex editing capability with the endogenous

tRNA-processing system. Proc Natl Acad Sci, USA. 2015; 112: 3570–3575. https://doi.org/10.1073/

pnas.1420294112 PMID: 25733849

66. Liu W, Xie X, Ma X, Li J, Chen J, Liu YG. DSDecode: A Web-based tool for decoding of sequencing

chromatograms for genotyping of targeted mutations. Mol Plant. 2015; 8: 1431–1433. https://doi.org/

10.1016/j.molp.2015.05.009 PMID: 26032088

67. Dehairs J, Talebi A, Cherifi Y, Swinnen JV. CRISP-ID: decoding CRISPR mediated indels by Sanger

sequencing. Scientific Rep. 2016; 6: 28973. https://doi.org/10.1038/srep28973 PMID: 27363488

68. Phung NTP, Mai CD, Mournet P, Frouin J, Droc G, Ta NK, et al. Characterization of a panel of Vietnam-

ese rice varieties using DArT and SNP markers for association mapping purposes. BMC Plant Biol.

2014; 14: 497. https://doi.org/10.1186/s12870-014-0371-7 PMID: 25524444

69. Kumar S, Stecher G, Li M, Knyaz C, Tamura K. MEGA X: Molecular evolutionary genetics analysis

across computing platforms. Mol Biol Evol. 2018; 35: 1547–1549. https://doi.org/10.1093/molbev/

msy096 PMID: 29722887

PLOS GENETICS A panicle related QTL associated with ANK and ANK-TPR genes

PLOS Genetics | https://doi.org/10.1371/journal.pgen.1009594 June 7, 2021 24 / 24